File size: 13,896 Bytes
eef4ddc |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 |
# ============================================================
# Unit 6_upload.py - 智能上传(优先使用最佳模型)
# ============================================================
import gymnasium as gym
import panda_gym
import numpy as np
import os
import shutil
from stable_baselines3 import A2C
from stable_baselines3.common.env_util import make_vec_env
from stable_baselines3.common.vec_env import VecNormalize, VecVideoRecorder
from stable_baselines3.common.evaluation import evaluate_policy
from huggingface_hub import HfApi, create_repo
# ============================================================
# 配置参数(⚠️ 修改这里)
# ============================================================
USERNAME = "ImaghT" # 你的 HF 用户名
MODEL_NAME = "a2c-PandaReachDense-v3"
ENV_ID = "PandaReachDense-v3"
N_EVAL_EPISODES = 20 # 增加评估episodes获得更准确结果
repo_id = f"{USERNAME}/{MODEL_NAME}"
# ============================================================
# 1. 智能文件检测(优先使用最佳模型)
# ============================================================
print("="*60)
print("🔍 检测可用模型文件...")
print("="*60)
# 文件路径定义
BEST_MODEL_PATH = "/home/eason/Workspace/RL/Unit_6/logs/best_model"
BEST_VEC_NORMALIZE_PATH = "/home/eason/Workspace/RL/Unit_6/logs/best_model_vecnormalize.pkl"
FINAL_MODEL_PATH = "a2c-PandaReachDense-v3"
FINAL_VEC_NORMALIZE_PATH = "vec_normalize.pkl"
# 🎯 优先级检查:最佳模型 > 最终模型
if os.path.exists(f"{BEST_MODEL_PATH}.zip") and os.path.exists(BEST_VEC_NORMALIZE_PATH):
print("✅ 发现训练期间保存的最佳模型(推荐使用)")
MODEL_PATH = BEST_MODEL_PATH
VEC_NORMALIZE_PATH = BEST_VEC_NORMALIZE_PATH
model_source = "best_model"
elif os.path.exists(f"{FINAL_MODEL_PATH}.zip") and os.path.exists(FINAL_VEC_NORMALIZE_PATH):
print("✅ 发现最终训练模型")
MODEL_PATH = FINAL_MODEL_PATH
VEC_NORMALIZE_PATH = FINAL_VEC_NORMALIZE_PATH
model_source = "final_model"
else:
print("❌ 错误: 未找到可用的模型文件!")
print("\n请确保以下文件之一存在:")
print(f" 方案1: {BEST_MODEL_PATH}.zip + {BEST_VEC_NORMALIZE_PATH}")
print(f" 方案2: {FINAL_MODEL_PATH}.zip + {FINAL_VEC_NORMALIZE_PATH}")
print("\n请先运行 Unit 6.py 训练代码。")
exit(1)
print(f"📁 使用模型: {MODEL_PATH}")
print(f"📁 使用归一化: {VEC_NORMALIZE_PATH}")
print(f"📊 模型来源: {model_source}\n")
# ============================================================
# 2. 加载模型
# ============================================================
print("加载模型...")
eval_env = make_vec_env(ENV_ID, n_envs=1)
eval_env = VecNormalize.load(VEC_NORMALIZE_PATH, eval_env)
eval_env.training = False
eval_env.norm_reward = False
model = A2C.load(MODEL_PATH, env=eval_env)
print("✅ 模型加载成功\n")
# ============================================================
# 3. 评估模型
# ============================================================
print("="*60)
print(f"🧪 开始评估 ({N_EVAL_EPISODES} episodes)...")
print("="*60)
mean_reward, std_reward = evaluate_policy(
model,
eval_env,
n_eval_episodes=N_EVAL_EPISODES,
deterministic=True
)
score = mean_reward - std_reward
print("\n" + "="*60)
print("📊 评估结果:")
print(f" Mean Reward: {mean_reward:.2f}")
print(f" Std Reward: {std_reward:.2f}")
print(f" Score (mean-std): {score:.2f}")
print(f" 通过基准线: -3.5")
if score >= -3.5:
print(f" ✅ 状态: PASSED")
status_emoji = "✅"
else:
print(f" ❌ 状态: NOT PASSED (还差 {-3.5 - score:.2f} 分)")
status_emoji = "❌"
print("="*60 + "\n")
# ============================================================
# 4. 生成演示视频
# ============================================================
print("🎬 生成演示视频...")
video_folder = "/home/eason/Workspace/RL/Unit_6/video_upload"
os.makedirs(video_folder, exist_ok=True)
video_env = make_vec_env(ENV_ID, n_envs=1)
video_env = VecNormalize.load(VEC_NORMALIZE_PATH, video_env)
video_env.training = False
video_env.norm_reward = False
video_env = VecVideoRecorder(
video_env,
video_folder,
record_video_trigger=lambda x: x == 0,
video_length=500,
name_prefix="panda-reach-agent"
)
obs = video_env.reset()
for _ in range(500):
action, _ = model.predict(obs, deterministic=True)
obs, _, _, _ = video_env.step(action)
video_env.close()
print(f"✅ 视频已生成\n")
# ============================================================
# 5. 检查训练日志(可选信息)
# ============================================================
training_info = ""
if os.path.exists("/home/eason/Workspace/RL/Unit_6/logs/evaluations.npz"):
try:
evaluations = np.load("/home/eason/Workspace/RL/Unit_6/logs/evaluations.npz")
timesteps = evaluations['timesteps']
results = evaluations['results']
# 获取训练过程信息
total_evals = len(timesteps)
final_timestep = timesteps[-1] if len(timesteps) > 0 else "Unknown"
best_eval_reward = np.max(results.mean(axis=1)) if len(results) > 0 else "Unknown"
training_info = f"""
## Training Monitoring
This model was trained with comprehensive monitoring:
- **Total Evaluations**: {total_evals} (every 500,000 steps)
- **Final Training Step**: {final_timestep:,}
- **Best Evaluation Reward**: {best_eval_reward:.2f}
- **Model Source**: {"Best model from training" if model_source == "best_model" else "Final training model"}
- **Callbacks Used**: EvalCallback, CheckpointCallback
- **TensorBoard Logging**: Enabled
"""
print(f"📈 发现训练日志: {total_evals} 次评估记录")
except Exception as e:
print(f"⚠️ 读取训练日志失败: {e}")
training_info = "\n## Training Monitoring\n\nModel trained with monitoring callbacks.\n"
else:
training_info = "\n## Training Configuration\n\nStandard training without detailed monitoring.\n"
# ============================================================
# 6. 创建增强版 README.md
# ============================================================
readme_content = f"""---
library_name: stable-baselines3
tags:
- PandaReachDense-v3
- deep-reinforcement-learning
- reinforcement-learning
- stable-baselines3
- robotics
- panda-gym
model-index:
- name: A2C
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: PandaReachDense-v3
type: PandaReachDense-v3
metrics:
- type: mean_reward
value: {mean_reward:.2f} +/- {std_reward:.2f}
name: mean_reward
verified: false
---
# {status_emoji} **A2C** Agent playing **PandaReachDense-v3**
This is a trained model of a **A2C** agent playing **PandaReachDense-v3**
using the [stable-baselines3 library](https://github.com/DLR-RM/stable-baselines3)
and the [Deep Reinforcement Learning Course](https://huggingface.co/deep-rl-course/unit6).
This environment is part of the [Panda-Gym](https://github.com/qgallouedec/panda-gym) environments and includes robotic manipulation tasks where the robot arm needs to reach a target position.
## 🏆 Evaluation Results
| Metric | Value |
|--------|-------|
| Mean Reward | {mean_reward:.2f} |
| Std Reward | {std_reward:.2f} |
| **Score (mean - std)** | **{score:.2f}** |
| Baseline Required | -3.5 |
| Evaluation Episodes | {N_EVAL_EPISODES} |
| Status | {status_emoji} {"**PASSED**" if score >= -3.5 else "**NOT PASSED**"} |
| Model Source | {model_source.replace('_', ' ').title()} |
{training_info}
## 🚀 Usage
```python
import gymnasium as gym
import panda_gym
from stable_baselines3 import A2C
from stable_baselines3.common.env_util import make_vec_env
from stable_baselines3.common.vec_env import VecNormalize
# Load environment and normalization
env = make_vec_env("PandaReachDense-v3", n_envs=1)
env = VecNormalize.load("vec_normalize.pkl", env)
# ⚠️ CRITICAL: disable training mode and reward normalization at test time
env.training = False
env.norm_reward = False
# Load model
model = A2C.load("a2c-PandaReachDense-v3", env=env)
# Run inference
obs = env.reset()
for _ in range(1000):
action, _states = model.predict(obs, deterministic=True)
obs, reward, done, info = env.step(action)
if done:
obs = env.reset()
```
## 🔧 Training Configuration
- **Algorithm**: A2C (Advantage Actor-Critic)
- **Policy**: MultiInputPolicy (for Dict observation spaces)
- **Environment**: PandaReachDense-v3
- **Total Timesteps**: 200,0000
- **Number of Parallel Envs**: 64
- **Normalization**: VecNormalize (observation + reward)
- **Observation Clipping**: 10.0
- **Evaluation Frequency**: Every 500,000 steps
- **Checkpoint Frequency**: Every 500,000 steps
## 🤖 Model Architecture
The agent uses a **MultiInputPolicy** because the observation space is a dictionary containing:
- `observation`: Robot joint positions, velocities, and gripper state
- `desired_goal`: Target position coordinates (x, y, z)
- `achieved_goal`: Current end-effector position coordinates (x, y, z)
The goal is to minimize the distance between `achieved_goal` and `desired_goal`.
## 📈 Performance Notes
- **Reward Range**: Typically from -50 (far from target) to 0 (at target)
- **Success Criteria**: Achieving mean reward > -3.5 consistently
- **Episode Length**: Usually 50 steps per episode
- **Convergence**: Expect improvement after 200k-500k steps
## 🎯 Tips for Reproduction
1. **Normalization is Critical**: Always use VecNormalize for robotic tasks
2. **MultiInputPolicy Required**: Dict observation spaces need special handling
3. **Sufficient Training**: 1M+ timesteps recommended for stable performance
4. **Evaluation**: Use deterministic=True for consistent evaluation results
"""
# ============================================================
# 7. 准备上传文件
# ============================================================
print("📦 准备上传文件...")
upload_folder = "/home/eason/Workspace/RL/Unit_6/upload_temp"
os.makedirs(upload_folder, exist_ok=True)
# 保存 README
readme_path = os.path.join(upload_folder, "README.md")
with open(readme_path, "w", encoding="utf-8") as f:
f.write(readme_content)
print(f"✅ 创建 README.md")
# 复制模型文件(重命名为标准名称)
model_dest = os.path.join(upload_folder, f"{MODEL_NAME}.zip")
shutil.copy(f"{MODEL_PATH}.zip", model_dest)
print(f"✅ 复制模型文件: {MODEL_PATH}.zip -> {MODEL_NAME}.zip")
# 复制归一化文件(重命名为标准名称)
vec_norm_dest = os.path.join(upload_folder, "vec_normalize.pkl")
shutil.copy(VEC_NORMALIZE_PATH, vec_norm_dest)
print(f"✅ 复制归一化文件: {VEC_NORMALIZE_PATH} -> vec_normalize.pkl")
# 复制视频文件
video_files = [f for f in os.listdir(video_folder) if f.endswith(".mp4")]
if video_files:
video_src = os.path.join(video_folder, video_files[0])
video_dest = os.path.join(upload_folder, "replay.mp4")
shutil.copy(video_src, video_dest)
print(f"✅ 复制视频文件")
else:
print(f"⚠️ 未找到视频文件(可选)")
# 可选:复制训练日志
if os.path.exists("/home/eason/Workspace/RL/Unit_6/logs/evaluations.npz"):
eval_dest = os.path.join(upload_folder, "training_evaluations.npz")
shutil.copy("/home/eason/Workspace/RL/Unit_6/logs/evaluations.npz", eval_dest)
print(f"✅ 复制训练评估日志")
# ============================================================
# 8. 上传到 Hugging Face Hub
# ============================================================
print(f"\n🚀 上传到 {repo_id}...")
api = HfApi()
try:
# 创建仓库(如果已存在则跳过)
create_repo(repo_id, repo_type="model", exist_ok=True)
print(f"✅ 仓库已创建/验证")
except Exception as e:
print(f"⚠️ 仓库警告: {e}")
try:
# 上传整个文件夹
commit_message = f"A2C PandaReach ({model_source}) - Mean: {mean_reward:.2f}, Std: {std_reward:.2f}, Score: {score:.2f}"
api.upload_folder(
folder_path=upload_folder,
repo_id=repo_id,
repo_type="model",
commit_message=commit_message
)
print(f"\n{'='*60}")
print("🎉 上传成功!")
print(f"{'='*60}")
print(f"🔗 模型页面: https://huggingface.co/{repo_id}")
print(f"🏆 检查进度: https://huggingface.co/spaces/ThomasSimonini/Check-my-progress-Deep-RL-Course")
print(f"📊 模型来源: {model_source.replace('_', ' ').title()}")
print(f"🎯 评估分数: {score:.2f} ({'通过' if score >= -3.5 else '未通过'})")
print(f"{'='*60}\n")
except Exception as e:
print(f"\n❌ 上传失败: {e}")
print(" 请检查:")
print(" 1. 是否已运行 'huggingface-cli login'")
print(" 2. 网络连接是否正常")
print(" 3. 用户名是否正确\n")
finally:
# 清理临时文件
shutil.rmtree(upload_folder)
print("🧹 清理临时文件")
print("✨ 完成!")
# ============================================================
# 9. 额外信息输出
# ============================================================
print("\n" + "="*60)
print("📋 上传总结")
print("="*60)
print(f"📁 上传的文件:")
print(f" - {MODEL_NAME}.zip (模型)")
print(f" - vec_normalize.pkl (归一化参数)")
print(f" - README.md (文档)")
print(f" - replay.mp4 (演示视频)")
if os.path.exists("/home/eason/Workspace/RL/Unit_6/logs/evaluations.npz"):
print(f" - training_evaluations.npz (训练日志)")
print(f"\n🎯 关键信息:")
print(f" - 使用了 {'最佳' if model_source == 'best_model' else '最终'} 模型")
print(f" - 评估分数: {score:.2f}")
print(f" - 状态: {'✅ 通过' if score >= -3.5 else '❌ 未通过'}")
print("="*60) |